SOCIETY BLOG
In This Section

De-mystifying Artificial Intelligence for Audit


Nov 13, 2019
Artificial intelligence (AI) is all around us—you have probably interacted with it today while using your phone, browsing the internet or chatting with phone support. Whether we realize it or not, we have already embraced AI as a tool to make life easier and more efficient. The audit industry, like other established professions, is relatively behind when it comes to adopting the technology.

To understand whether AI is right for your firm, it’s critical to hear from those who are putting it into practice, to know why it’s important and how it helps auditors.

AI-Powered

ARTIFICIAL INTELLIGENCE 101

AI is the implementation of a machine or a computer program to think and learn. Over 60 years ago, the first AI projects focused on tasks such as language translation, with the U.S. government funding a project to determine whether a machine could translate between English and Russian. Due to the limited computer power of the day, progress was limited. More recent advances, such as IBM Watson and smartphone assistants, have brought us closer to thinking computers.

AI and the related technology of machine learning have come to automate many tasks that were previously done manually and introduced the ability to analyze 100 percent of a dataset. More importantly, these technologies offer the ability to test data against controls that go beyond rules and into true anomaly detection. For auditors, this means no human is required to create tests, write scripts or remember all the procedural steps. It also means auditors do not have to rely on random sampling to provide reasonable assurance to clients.

By analyzing the entirety of the general ledger and reporting based on risk, rather than sampling, AI flags transactions for investigation based on how they deviate from the total data set, such as unusual payments or activities that would not normally be caught by traditional testing practices.

Will AI replace accountants? The answer is definitively “no.” It cannot replace the experience and judgement of auditors, nor can it understand and manage the relationships between firms and clients. AI works alongside people, automating and accelerating large and complex risk analysis, and it assists with decision making when it comes to identifying misstatements and providing guidance to clients.

AI IN PRACTICE

Top 60 firm Blue & Co. believes in building strong, long-term relationships with clients and following the philosophy of making your clients your friends and your friends your clients. The firm first looked into AI as a way to keep up with the way the business world was moving, including the increased use of sophisticated technology and larger data sets by clients.

“We’re counting on AI to reduce a lot of the repetitive steps we do now, reducing what normally takes days into a matter of hours or minutes,” says Mark Wischmeier, CPA and Senior Manager at Blue & Co. “Let’s say you’re looking at an employee benefit plan and checking what contributions were made throughout the whole year. Instead of taking a small sample and thinking that’s reasonable, AI lets you test 100 percent of the population. You can assess the entire data set in minutes instead of having an auditor spend hours or days looking at a tiny percentage only.”
"AI will eventually become like Excel is now. We won't need to know how the technology works in detail just like we don't need to know how the engines in our cars work. As long as you make yourself ready to know how to use AI, you will be good to go."

In addition to saving time, the expanded coverage of AI allows firms to go back to the client with more information and insights. Says Wischmeier, “It’s good to have in-depth discussions with clients to understand their needs in different ways. Ask them what their problems are, where they struggle month to month, and come up with ideas on how to use AI and data analytics differently to help them in the most effective manner possible.” He also believes that even though clients may not be familiar with AI, their exposure to other types of data-driven tools will make it an easier sell. “I’ve seen some terrific things that clients have done with analytics platforms. The success they’ve had with these should give them the confidence they need to help make AI a more natural extension that they understand.”

As important as AI is to a firm’s technology stack, Blue & Co. believes the people must also adapt to the change. “We know this is the way the industry is moving and, right from the top, we’re committed to bringing new technologies into the firm, including a focus on staffing. When you start moving outside the traditional norm, you need to go beyond traditional CPA skillsets and consider financial analysts, coders and scriptwriters, and data analytics people.”

As the practice continues their journey towards AI, Wischmeier has advice for others curious about the technology. “I’d recommend every firm keep tabs on what’s out there in terms of AI, bots and machine learning. It’s changing almost daily so it pays to keep up and understand the potential value added to clients.”

Big-Data-Powered

AI FOR THE NEXT GENERATION OF AUDITORS

Dr. Sung-Jin Park, assistant professor in accounting at Indiana University South Bend, believes early stage exposure to artificial intelligence is critical in preparing students for the profession of the future—a future that’s not as far off as some would think.

“The entire business world, including audit, is changing,” says Park. “Big Data initiated the current trend, starting with analytics, then machine learning, and now to AI. It’s conceivable that going forward, statistical sampling in audit practices will become irrelevant and AI will be a necessity. Students must get exposure to these technologies now and that’s why we’re trying to include it in our curriculum.”

Educating the next generation of auditors means following where the industry is going while also not going too deep into the complexities of the software. “We’re not trying to train programmers, rather we’re exposing today’s students to AI to understand how it works in accounting practices and prepare them for long-running careers. While it might be somewhat premature for full-fledged adoption in the academic world, accounting associations are already making a big push for the technology and eventually entire firms will be on board. The deeper AI training will probably occur in the firms but at the very least, our courses must familiarize students before they enter the workforce.”

As a researcher of audit pricing and market structures, Park has a unique view on the potential impacts of AI. “Pricing is all about the auditor’s assessment of risk, which comes from their verification work, and the actual procedures and tests are often based on the samples they collect,” says Park. “In the past, evaluating the data representing all the transactions was virtually impossible because no one had unlimited amounts of time to go through it. Now, with auditors having AI and machine learning go through far more transactions than sampling could, I expect more confident assessments, which means a reduction in detection risk. This could result in higher quality financial reporting for clients and the possibility of lowering audit costs.”

With all the news and fears around AI replacing human jobs, the question comes up of whether students are motivated to take courses including the technology. “I’m pretty sure students want to learn AI and, more importantly, how to use data more efficiently. Analytics training is already a big component of our curriculum and this will eventually lead to machine learning and AI. They always hear about the importance of data in journals and media every day and are highly motivated to know more about it. Even younger professors are looking to take retraining in order to incorporate AI into more of their classes, particularly for audit.”

While AI is still new, Park believes the technology will become commonplace in the near future. “A popular misconception among the public is that AI is too complicated to understand or difficult to use in practice,” he says. “From my viewpoint, AI will eventually become like Excel is now. We won’t need to know how the technology works in detail just like we don’t need to know how the engines in our cars work. As long as you make yourself ready to know how to use AI, you will be good to go.”


Leave a comment